TUTORIALS

HiPC 2006 Tutorials Flyer (PDF)

• Tutorial I       CANCELLED
  Getting Going with the Grid and Its Applications
  Presenter: Dr. Mark Baker
  Professor, School of Systems Engineering, The University of Reading, UK
  Full Day tutorial
  

  Abstract
  The Grid concepts and ideas have moved away from the world of joining high-performance computers together to more generalised systems based on service oriented architectures that can provide infrastructure for a range of applications. In this tutorial the basic concepts of the Grid are explained, including its core components, current instantiations such as Globus, and the description of a range of representative applications. The tutorial will also have an overview of Project Garuda, which is a effort to Grid enable multiple computing centres in India and is being led by CDAC. The tutorial will include some hands on work-session.

   

• Tutorial II
  MPI - Portable Scalable Programming for High Performance Computing
  Presenter: Ewing ("Rusty") Lusk
  Senior Computer Scientist, Acting Division Director, Mathematics and Computer Science Division, Argonne National Labs
  Half Day tutorial
  

  Abstract
  MPI (Message Passing Interface) has been the dominant standard on which parallel applications have been developed for the last decade. This tutorial will provide insight into the fundamental concepts and the key constructs of MPI and how they enable rapid development of parallel applications. The tutorial will drill down on the MPI-2 standard as well with Dynamic Process Creation to grow/shrink the size of a parallel application; MPI-One sided communication; and MPI-IO. The tutorial will also give an overview of the fundamentals of debugging and performance tuning parallel applications.

   

• Tutorial III
  Cluster and Parallel File systems for High Performance Computing Clusters
  Presenter: Dr. Gautam Shah
  Senior Technical Staff Member, IBM Systems Development Group
  Half Day tutorial
  

  Abstract
  Cluster and Parallel File systems are fast becoming a very important component of clusters and parallel systems. In this Tutorial we will discuss a) the fundamental concepts that define Cluster and Parallel File systems; b) the basic building blocks that make up cluster/parallel file systems; c) overview of the performance choices and their tradeoffs in designing cluster and parallel file systems; d) a comparative survey of the state of the art of existing cluster and parallel file systems. e) future directions for cluster and parallel file systems; and a detailed overview of GPFS - its architecture, design, and performance as a reference cluster/parallel file system.

Please visit later for more information on these tutorials.